.ipynb .pdf

repository open issue

Binder

Contents

8. Overlapping densities

8.1. 概要

8.2. Plotlyによる作図方法

8.3. MADB Labを用いた作図例

8.3.1. 下準備

import pandas as pd
import plotly.figure_factory as ff

import warnings
warnings.filterwarnings('ignore')

# 前処理の結果，以下に分析対象ファイルが格納されていることを想定
PATH_DATA = '../../data/preprocess/out/magazines.csv'
# Jupyter Book用のPlotlyのrenderer
RENDERER = 'plotly_mimetype+notebook'

def show_fig(fig):
    """Jupyter Bookでも表示可能なようRendererを指定"""
    fig.show(renderer=RENDERER)

df = pd.read_csv(PATH_DATA)

8.3.2. 雑誌別・作品別の合計連載週数

df_plot = \
    df.value_counts(['mcname', 'cname']).reset_index(name='weeks')

df_plot = \
    df.value_counts(['mcname', 'cname']).reset_index(name='weeks')
# distplot用に中間集計
mcnames = df_plot.mcname.unique()
hist_data = [
    df_plot[df_plot['mcname']==mc]['weeks'].values.reshape(-1)
    for mc in mcnames]
fig = ff.create_distplot(
    hist_data, mcnames, show_hist=False)
fig.update_layout(title_text='雑誌別・作品別の合計連載週数')
show_fig(fig)

fig = ff.create_distplot(
    hist_data, mcnames, show_hist=False)
fig.update_layout(title_text='雑誌別・作品別の合計連載週数')
fig.update_xaxes(range=[0, 200])
show_fig(fig)

8.3.3. 雑誌別・作者別の合計連載週数

df_plot = \
    df.value_counts(['mcname', 'creator']).reset_index(name='weeks')
# distplot用に中間集計
mcnames = df_plot.mcname.unique()
hist_data = [
    df_plot[df_plot['mcname']==mc]['weeks'].values.reshape(-1)
    for mc in mcnames]
fig = ff.create_distplot(
    hist_data, mcnames, show_hist=False)
fig.update_layout(title_text='雑誌別・作者別の合計連載週数')
show_fig(fig)

fig = ff.create_distplot(
    hist_data, mcnames, show_hist=False)
fig.update_layout(title_text='雑誌別・作者別の合計連載週数')
fig.update_xaxes(range=[0, 200])
show_fig(fig)

previous

7. Stacked histograms

next

9. Ridgeline plot